Re: [core] draft-ietf-core-observe-05 - "obs"

Zach Shelby <> Fri, 20 April 2012 15:38 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id F2ACE21F8731 for <>; Fri, 20 Apr 2012 08:38:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -3.599
X-Spam-Status: No, score=-3.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id 8kxroGXxPG8n for <>; Fri, 20 Apr 2012 08:38:44 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id C27E821F85E1 for <>; Fri, 20 Apr 2012 08:38:43 -0700 (PDT)
Received: from [] ( []) (authenticated bits=0) by (8.13.8/8.13.4) with ESMTP id q3KFccF1018290; Fri, 20 Apr 2012 18:38:39 +0300
Mime-Version: 1.0 (Apple Message framework v1084)
Content-Type: text/plain; charset=us-ascii
From: Zach Shelby <>
In-Reply-To: <>
Date: Fri, 20 Apr 2012 18:38:37 +0300
Content-Transfer-Encoding: quoted-printable
Message-Id: <>
References: <>
To: Klaus Hartke <>
X-Mailer: Apple Mail (2.1084)
Subject: Re: [core] draft-ietf-core-observe-05 - "obs"
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: "Constrained RESTful Environments \(CoRE\) Working Group list" <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Fri, 20 Apr 2012 15:38:45 -0000


I would like to keep the 'obs' link attribute. From the experience deploying products and systems using CoAP so far, we have actually found this attribute useful. Two reasons:

1. Although in theory the protocol still works through trial and error as you describe below, it is useful for a CoAP client to plan its behavior before making a request.

2. Knowledge of the Observability of a resource is useful at the application layer, where some M2M application needs to make use of these resources. The behavior of dealing with a polling-only resource vs. an observable resource is pretty different, and an observable resource might result in a different graphical representation to the user which needs to be known already during discovery.

I like the philosophy otherwise for finding things to take away, but not this one please.


On Apr 17, 2012, at 3:14 AM, Klaus Hartke wrote:

> In the spirit of "In protocol design, perfection has been reached not
> when there is nothing left to add, but when there is nothing left to
> take away" [1]:
> Do we really need the "obs" link attribute?
> Some thoughts:
> * If a client wants to have a fresh representation of a resource over
> a period of time, it can include the Observe option in its request. If
> the server does not support -observe, the client can poll the resource
> to achieve its goal.
> * If a client for whatever reason only wants to have a fresh
> representation of a resource over a period of time if the server
> supports -observe, it can include Observe option in its request and
> not poll if the if the server does not support -observe.
> * If a client wants a single snapshot representation of a resource, it
> can omit the Observe option from its request.
> Under what circumstances can a client not be sure if it wants to have
> a fresh representation of a resource over a period of time, so a hint
> from the server is needed?
> Klaus
> [1]
> _______________________________________________
> core mailing list

Zach Shelby, Chief Nerd, Sensinode Ltd.  - My blog "On the Internet of Things" - My book "6LoWPAN: The Wireless Embedded Internet"
Mobile: +358 40 7796297